idnits 2.17.1 draft-ietf-ipo-impairments-03.txt: -(160): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(193): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(206): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(209): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(217): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(909): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(920): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(1031): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(1078): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(1079): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(1098): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding -(1124): Line appears to be too long, but this could be caused by non-ascii characters in UTF-8 encoding Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- ** Looks like you're using RFC 2026 boilerplate. This must be updated to follow RFC 3978/3979, as updated by RFC 4748. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- ** Missing document type: Expected "INTERNET-DRAFT" in the upper left hand corner of the first page ** The document seems to lack a 1id_guidelines paragraph about Internet-Drafts being working documents. ** The document seems to lack a 1id_guidelines paragraph about 6 months document validity -- however, there's a paragraph with a matching beginning. Boilerplate error? ** The document is more than 15 pages and seems to lack a Table of Contents. == There are 20 instances of lines with non-ascii characters in the document. == No 'Intended status' indicated for this document; assuming Proposed Standard Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- ** The document seems to lack an IANA Considerations section. (See Section 2.2 of https://www.ietf.org/id-info/checklist for how to handle the case when there are no actions for IANA.) ** The document seems to lack an Authors' Addresses Section. ** There are 493 instances of too long lines in the document, the longest one being 5 characters in excess of 72. Miscellaneous warnings: ---------------------------------------------------------------------------- -- The document seems to lack a disclaimer for pre-RFC5378 work, but may have content which was first submitted before 10 November 2008. If you have contacted all the original authors and they are all willing to grant the BCP78 rights to the IETF Trust, then this is fine, and you can ignore this comment. If not, you may need to add the pre-RFC5378 disclaimer. (See the Legal Provisions document at https://trustee.ietf.org/license-info for more information.) -- The document date (September 2002) is 7892 days in the past. Is this intentional? Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) == Missing Reference: 'GMPLS' is mentioned on line 45, but not defined == Missing Reference: 'Gerstel2000' is mentioned on line 69, but not defined == Missing Reference: 'Strand2001' is mentioned on line 69, but not defined == Unused Reference: 'Bra96' is defined on line 1191, but no explicit reference was found in the text == Unused Reference: 'CBD00' is defined on line 1194, but no explicit reference was found in the text == Unused Reference: 'Doverspike00' is defined on line 1201, but no explicit reference was found in the text == Unused Reference: 'Gerstel 2000' is defined on line 1205, but no explicit reference was found in the text == Unused Reference: 'KRB01a' is defined on line 1209, but no explicit reference was found in the text == Unused Reference: 'KRB01b' is defined on line 1213, but no explicit reference was found in the text == Unused Reference: 'Moy98' is defined on line 1217, but no explicit reference was found in the text -- Possible downref: Normative reference to a draft: ref. 'Chaudhuri00' -- Possible downref: Non-RFC (?) normative reference: ref. 'Goldstein94' -- Possible downref: Non-RFC (?) normative reference: ref. 'ITU' -- Possible downref: Non-RFC (?) normative reference: ref. 'Kaminow97' == Outdated reference: A later version (-02) exists of draft-chiu-strand-unique-olcp-01 -- Possible downref: Normative reference to a draft: ref. 'Strand01' -- Possible downref: Non-RFC (?) normative reference: ref. 'Strand01b' -- Possible downref: Non-RFC (?) normative reference: ref. 'Yates99' == Outdated reference: A later version (-03) exists of draft-awduche-mpls-te-optical-01 -- No information found for draft-ietf-gmpls- - is the name correct? -- No information found for draft-ospf-generalized- - is the name correct? Summary: 8 errors (**), 0 flaws (~~), 14 warnings (==), 11 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Internet Draft John Strand (Editor) 3 Document: draft-ietf-ipo-impairments-03.txt AT&T 4 Informational Track 5 Expiration Date: March 2003 Angela Chiu (Editor) 6 Celion Networks 8 September 2002 10 Impairments And Other Constraints On Optical Layer Routing 12 Status of this Memo 14 This document is an Internet-Draft and is in full conformance with 15 all provisions of Section 10 of RFC2026. Internet-Drafts are 16 working documents of the Internet Engineering Task Force (IETF), its 17 areas, and its working groups. Note that other groups may also 18 distribute working documents as Internet-Drafts. 20 Internet-Drafts are draft documents valid for a maximum of six 21 months and may be updated, replaced, or made obsolete by other 22 documents at any time. It is inappropriate to use Internet-Drafts as 23 reference material or to cite them other than as "work in progress." 25 The list of current Internet-Drafts can be accessed at 26 http://www.ietf.org/ietf/1id-abstracts.txt. 27 The list of Internet-Draft Shadow Directories can be accessed at 28 http://www.ietf.org/shadow.html. 30 Abstract 31 Optical networking poses a number challenges for GMPLS. Optical 32 technology is fundamentally an analog rather than digital technology; 33 and the optical layer is lowest in the transport hierarchy and hence 34 has an intimate relationship with the physical geography of the 35 network. This contribution surveys some of the aspects of optical 36 networks which impact routing and identifies possible GMPLS responses 37 for each: (1) Constraints arising from the design of new software 38 controllable network elements, (2) Constraints in a single all- 39 optical domain without wavelength conversion, (3) Complications 40 arising in more complex networks incorporating both all-optical and 41 opaque architectures, and (4) Impacts of diversity constraints. 43 1. Introduction 45 GMPLS [GMPLS] aims to extend MPLS to encompass a number of transport 46 architectures. Included are optical networks incorporating a number 47 of all-optical and opto-electronic elements such as optical cross- 48 connects with both optical and electrical fabrics, transponders, and 49 optical add-drop multiplexers. Optical networking poses a number 50 challenges for GMPLS. Optical technology is fundamentally an analog 51 Impairments And Other Constraints September 2002 52 On Optical Layer Routing 54 rather than digital technology; and the optical layer is lowest in 55 the transport hierarchy and hence has an intimate relationship with 56 the physical geography of the network. 58 GMPLS already has incorporated extensions to deal with some of the 59 unique aspects of the optical layer. This contribution surveys some 60 of the aspects of optical networks which impact routing and 61 identifies possible GMPLS responses for each. Routing constraints 62 and/or complications arising from the design of network elements, 63 the accumulation of signal impairments, and from the need to 64 guarantee the physical diversity of some circuits are discussed. 66 Since the purpose of this draft is to further the specification of 67 GMPLS, alternative approaches to controlling an optical network are 68 not discussed. For discussions of some broader issues, see 69 [Gerstel2000] and [Strand2001]. 71 The organization of the contribution is as follows: 73 - Section 2 is a section requested by the sub-IP Area management 74 for all new drafts. It explains how this document fits into the 75 Area and into the IPO WG, and why it is appropriate for these 76 groups. 77 - Section 3 describes constraints arising from the design of new 78 software controllable network elements. 79 - Section 4 addresses the constraints in a single all-optical 80 domain without wavelength conversion. 81 - Section 5 extends the discussion to more complex networks. 82 incorporating both all-optical and opaque architectures. 83 - Section 6 discusses the impacts of diversity constraints. 84 - Section 7 deals with security requirements. 85 - Section 8 contains acknowledgments. 86 - Section 9 contains references. 87 - Section 10 contains contributing authors� addresses. 88 - Section 11 contains editors� addresses. 90 2. Sub-IP Area Summary And Justification Of Work 91 This draft merges and extends two previous drafts, draft-chiu- 92 strand-unique-olcp-02.txt and draft-banerjee-routing-impairments- 93 00.txt. These two drafts were made IPO working group documents to 94 form a basis for a design team at the Minneapolis meeting, where it 95 was also requested that they be merged to create a requirements 96 document for the WG. 98 In the larger sub-IP Area structure, this merged document describes 99 specific characteristics of optical technology and the requirements 100 they place on routing and path selection. It is appropriate for the 101 Impairments And Other Constraints September 2002 102 On Optical Layer Routing 104 IPO working group because the material is specific to optical 105 networks. It identifies and documents the characteristics of the 106 optical transport network that are important for selecting paths for 107 optical channels, which is a work area for the IPO WG. It is 108 appropriate work for this WG because the material covered is 109 directly aimed at establishing a framework and requirements for 110 routing in an optical network. 112 Related documents are: 113 draft-banerjee-routing-impairments-00.txt 114 draft-parent-obgp-01.txt 115 draft-bernstein-optical-bgp-00.txt 116 draft-hayata-ipo-carrier-needs-00.txt 117 draft-many-carrier-framework-uni-01.txt 118 draft-papadimitriou-ipo-non-linear-routing-impairm-01.txt 120 3. Reconfigurable Network Elements 122 3.1 Technology Background 124 Control plane architectural discussions (e.g., [Awduche99]) usually 125 assume that the only software reconfigurable network element is an 126 optical layer cross-connect (OLXC). There are however other 127 software reconfigurable elements on the horizon, specifically 128 tunable lasers and receivers and reconfigurable optical add-drop 129 multiplexers (OADM�s). These elements are illustrated in the 130 following simple example, which is modeled on announced Optical 131 Transport System (OTS) products: 132 + + 133 ---+---+ |\ /| +---+--- 134 ---| A |----|D| X Y |D|----| A |--- 135 ---+---+ |W| +--------+ +--------+ |W| +---+--- 136 : |D|-----| OADM |-----| OADM |-----|D| : 137 ---+---+ |M| +--------+ +--------+ |M| +---+--- 138 ---| A |----| | | | | | | |----| A |--- 139 ---+---+ |/ | | | | \| +---+--- 140 + +---+ +---+ +---+ +---+ + 141 D | A | | A | | A | | A | E 142 +---+ +---+ +---+ +---+ 143 | | | | | | | | 145 Figure 3-1: An OTS With OADM's - Functional Architecture 147 In Fig.3-1, the part that is on the inner side of all boxes labeled 148 "A" defines an all-optical subnetwork. From a routing perspective 149 two aspects are critical: 151 Impairments And Other Constraints September 2002 152 On Optical Layer Routing 154 - Adaptation: These are the functions done at the edges of the 155 subnetwork that transform the incoming optical channel into the 156 physical wavelength to be transported through the subnetwork. 157 - Connectivity: This defines which pairs of edge Adaptation 158 functions can be interconnected through the subnetwork. 160 In Fig. 3-1, D and E are DWDM�s and X and Y are OADM�s. The boxes 161 labeled "A" are adaptation functions. They map one or more input 162 optical channels assumed to be standard short reach signals into a 163 long reach (LR) wavelength or wavelength group which will pass 164 transparently to a distant adaptation function. Adaptation 165 functionality which affects routing includes: 166 - Multiplexing: Either electrical or optical TDM may be used to 167 combine the input channels into a single wavelength. This is 168 done to increase effective capacity: A typical DWDM might be 169 able to handle 100 2.5 Gb/sec signals (250 Gb/sec total) or 50 170 10 Gb/sec (500 Gb/sec total); combining the 2.5 Gb/sec signals 171 together thus effectively doubles capacity. After multiplexing 172 the combined signal must be routed as a group to the distant 173 adaptation function. 174 - Adaptation Grouping: In this technique, groups of k (e.g., 4) 175 wavelengths are managed as a group within the system and must be 176 added/dropped as a group. We will call such a group an 177 "adaptation grouping". Examples include so called "wave group" 178 and "waveband" [Passmore01]. Groupings on the same system may 179 differ in basics such as wavelength spacing, which constrain the 180 type of channels that can be accommodated. 181 - Laser Tunability: The lasers producing the LR wavelengths may 182 have a fixed frequency, may be tunable over a limited range, or 183 be tunable over the entire range of wavelengths supported by the 184 DWDM. Tunability speeds may also vary. 186 Connectivity between adaptation functions may also be limited: 187 - As pointed out above, TDM multiplexing and/or adaptation 188 grouping by the adaptation function forces groups of input 189 channels to be delivered together to the same distant adaptation 190 function. 191 - Only adaptation functions whose lasers/receivers are tunable to 192 compatible frequencies can be connected. 193 - The switching capability of the OADM�s may also be constrained. 194 For example: 195 o There may be some wavelengths that can not be dropped at 196 all. 197 o There may be a fixed relationship between the frequency 198 dropped and the physical port on the OADM to which it is 199 dropped. 200 o OADM physical design may put an upper bound on the number 201 of adaptation groupings dropped at any single OADM. 203 Impairments And Other Constraints September 2002 204 On Optical Layer Routing 206 For a fixed configuration of the OADM�s and adaptation functions 207 connectivity will be fixed: Each input port will essentially be 208 hard-wired to some specific distant port. However this connectivity 209 can be changed by changing the configurations of the OADM�s and 210 adaptation functions. For example, an additional adaptation grouping 211 might be dropped at an OADM or a tunable laser retuned. In each case 212 the port-to-port connectivity is changed. 214 These capabilities can be expected to be under software control. 215 Today the control would rest in the vendor-supplied Element 216 Management system (EMS), which in turn would be controlled by the 217 operator�s OS�s. However in principle the EMS could participate in 218 the GMPLS routing process. 220 3.2 Implications For Routing 222 An OTS of the sort discussed in Sec. 3.1 is essentially a 223 geographically distributed but blocking cross-connect system. The 224 specific port connectivity is dependent on the vendor design and 225 also on exactly what line cards have been deployed. 227 One way for GMPLS to deal with this architecture would be to view 228 the port connectivity as externally determined. In this case the 229 links known to GMPLS would be groups of identically routed 230 wavebands. If these were reconfigured by the external EMS the 231 resulting connectivity changes would need to be detected and 232 advertised within GMPLS. If the topology shown in Fig. 3-1 became a 233 tree or a mesh instead of the linear topology shown, the 234 connectivity changes could result in SRLG changes. 236 Alternatively, GMPLS could attempt to directly control this port 237 connectivity. The state information needed to do this is likely to 238 be voluminous and vendor specific. 240 4. Wavelength Routed All-Optical Networks 242 The optical networks presently being deployed may be called "opaque" 243 ([Tkach98]): each link is optically isolated by transponders doing 244 O/E/O conversions. They provide regeneration with retiming and 245 reshaping, also called 3R, which eliminates transparency to bit 246 rates and frame format. These transponders are quite expensive and 247 their lack of transparency also constrains the rapid introduction of 248 new services. Thus there are strong motivators to introduce 249 "domains of transparency" - all-optical subnetworks - larger than an 250 OTS. 252 The routing of lightpaths through an all-optical network has 253 received extensive attention. (See [Yates99] or [Ramaswami98]). 255 Impairments And Other Constraints September 2002 256 On Optical Layer Routing 258 When discussing routing in an all-optical network it is usually 259 assumed that all routes have adequate signal quality. This may be 260 ensured by limiting all-optical networks to subnetworks of limited 261 geographic size which are optically isolated from other parts of the 262 optical layer by transponders. This approach is very practical and 263 has been applied to date, e.g. when determining the maximum length 264 of an Optical Transport System (OTS). Furthermore operational 265 considerations like fault isolation also make limiting the size of 266 domains of transparency attractive. 268 There are however reasons to consider contained domains of 269 transparency in which not all routes have adequate signal quality. 270 From a demand perspective, maximum bit rates have rapidly increased 271 from DS3 to OC-192 and soon OC-768 (40 Gb/sec). As bit rates 272 increase it is necessary to increase power. This makes impairments 273 and nonlinearities more troublesome. From a supply perspective, 274 optical technology is advancing very rapidly, making ever-larger 275 domains possible. In this section we assume that these 276 considerations will lead to the deployment of a domain of 277 transparency that is too large to ensure that all potential routes 278 have adequate signal quality for all circuits. Our goal is to 279 understand the impacts of the various types of impairments in this 280 environment. 282 Note that as we describe later in the section there are many types 283 of physical impairments. Which of these need to be dealt with 284 explicitly when performing on-line distributed routing will vary 285 considerably and will depend on many variables, including: 286 - Equipment vendor design choices, 287 - Fiber characteristics, 288 - Service characteristics (e.g., circuit speeds), 289 - Network size, 290 - Network operator engineering and deployment strategies. 291 For example, a metropolitan network which does not intend to support 292 bit rates above 2.5 Gb/sec may not be constrained by any of these 293 impairments, while a continental or international network which 294 wished to minimize O/E/O regeneration investment and support 40 295 Gb/sec connections might have to explicitly consider many of them. 296 Also, a network operator may reduce or even eliminate their 297 constraint set by building a relatively small domain of transparency 298 to ensure that all the paths are feasible, or by using some 299 proprietary tools based on rules from the OTS vendor to pre-qualify 300 paths between node pairs and put them in a table that can be 301 accessed each time a routing decision has to be made through that 302 domain. 304 4.1 Problem Formulation 305 Impairments And Other Constraints September 2002 306 On Optical Layer Routing 308 We consider a single domain of transparency without wavelength 309 translation. Additionally due to the proprietary nature of DWDM 310 transmission technology, we assume that the domain is either single 311 vendor or architected using a single coherent design, particularly 312 with regard to the management of impairments. 314 We wish to route a unidirectional circuit from ingress client node X 315 to egress client node Y. At both X and Y, the circuit goes through 316 an O/E/O conversion which optically isolates the portion within our 317 domain. We assume that we know the bit rate of the circuit. Also, 318 we assume that the adaptation function at X may apply some Forward 319 Error Correction (FEC) method to the circuit. We also assume we know 320 the launch power of the laser at X. 322 Impairments can be classified into two categories, linear and 323 nonlinear. (See [Tkach98] for more on impairment constraints). 324 Linear effects are independent of signal power and affect 325 wavelengths individually. Amplifier spontaneous emission (ASE), 326 polarization mode dispersion (PMD), and chromatic dispersion are 327 examples. Nonlinearities are significantly more complex: they 328 generate not only impairments on each channel, but also crosstalk 329 between channels. 331 In the remainder of this section we first outline how two key linear 332 impairments (PMD and ASE) might be handled by a set of analytical 333 formulae as additional constraints on routing. We next discuss how 334 the remaining constraints might be approached. Finally we take a 335 broader perspective and discuss the implications of such constraints 336 on control plane architecture and also on broader constrained domain 337 of transparency architecture issues. 339 4.2 Polarization Mode Dispersion (PMD) 341 For a transparent fiber segment, the general PMD requirement is that 342 the time-average differential group delay (DGD) between two 343 orthogonal state of polarizations should be less than fraction a of 344 the bit duration, T=1/B, where B is the bit rate. The value of the 345 parameter a depends on three major factors: 1) margin allocated to 346 PMD, e.g. 1dB; 2) targeted outage probability, e.g. 4x10-5, and 3) 347 sensitivity of the receiver to DGD. A typical value for a is 10% 348 [ITU]. More aggressive designs to compensate for PMD may allow 349 values higher than 10%. (This would be a system parameter dependent 350 on the system design. It would need to be known to the routing 351 process.) 353 The PMD parameter (Dpmd) is measured in pico-seconds (ps) per 354 sqrt(km). The square of the PMD in a fiber span, denoted as span- 355 PMD-square is then given by the product of Dpmd**2 and the span 356 length. (A fiber span in a transparent network refers to a segment 357 between two optical amplifiers.) If Dpmd is constant, this results 358 Impairments And Other Constraints September 2002 359 On Optical Layer Routing 361 in a upper bound on the maximum length of an M-fiber-span 362 transparent segment, which is inversely proportional to the square 363 of the product of bit rate and Dpmd (the detailed equation is 364 omitted due to the format constraint - see [Strand01] for details). 366 For older fibers with a typical PMD parameter of 0.5 picoseconds per 367 square root of km, based on the constraint, the maximum length of 368 the transparent segment should not exceed 400km and 25km for bit 369 rates of 10Gb/s and 40Gb/s, respectively. Due to recent advances in 370 fiber technology, the PMD-limited distance has increased 371 dramatically. For newer fibers with a PMD parameter of 0.1 372 picosecond per square root of km, the maximum length of the 373 transparent segment (without PMD compensation) is limited to 10000km 374 and 625km for bit rates of 10Gb/s and 40Gb/, respectively. Still 375 lower values of PMD are attainable in commercially available fiber 376 today, and the PMD limit can be further extended if a larger value 377 of the parameter a (ratio of DGD to the bit period) can be 378 tolerated. In general, the PMD requirement is not an issue for most 379 types of fibers at 10Gb/s or lower bit rate. But it will become an 380 issue at bit rates of 40Gb/s and higher. 382 If the PMD parameter varies between spans, a slightly more 383 complicated equation results (see [Strand01]), but in any event the 384 only link dependent information needed by the routing algorithm is 385 the square of the link PMD, denoted as link-PMD-square. It is the 386 sum of the span-PMD-square of all spans on the link. 388 Note that when one has some viable PMD compensation devices and 389 deploy them ubiquitously on all routes with potential PMD issues in 390 the network, then the PMD constraint disappears from the routing 391 perspective. 393 4.3 Amplifier Spontaneous Emission 395 ASE degrades the optical signal to noise ratio (OSNR). An acceptable 396 optical SNR level (SNRmin) which depends on the bit rate, 397 transmitter-receiver technology (e.g., FEC), and margins allocated 398 for the impairments, needs to be maintained at the receiver. In 399 order to satisfy this requirement, vendors often provide some 400 general engineering rule in terms of maximum length of the 401 transparent segment and number of spans. For example, current 402 transmission systems are often limited to up to 6 spans each 80km 403 long. For larger transparent domains, more detailed OSNR 404 computations will be needed to determine whether the OSNR level 405 through a domain of transparency is acceptable. This would provide 406 flexibility in provisioning or restoring a lightpath through a 407 transparent subnetwork. 409 Assume that the average optical power launched at the transmitter is 410 P. The lightpath from the transmitter to the receiver goes through M 411 Impairments And Other Constraints September 2002 412 On Optical Layer Routing 414 optical amplifiers, with each introducing some noise power. Unity 415 gain can be used at all amplifier sites to maintain constant signal 416 power at the input of each span to minimize noise power and 417 nonlinearity. A constraint on the maximum number of spans can be 418 obtained [Kaminow97] which is proportional to P and inversely 419 proportional to SNRmin, optical bandwidth B, amplifier gain G-1 and 420 spontaneous emission factor n of the optical amplifier, assuming all 421 spans have identical gain and noise figure. (Again, the detailed 422 equation is omitted due to the format constraint - see [Strand01] 423 for details.) Let�s take a typical example. Assuming P=4dBm, 424 SNRmin=20dB with FEC, B=12.5GHz, n=2.5, G=25dB, based on the 425 constraint, the maximum number of spans is at most 10. However, if 426 FEC is not used and the requirement on SNRmin becomes 25dB, the 427 maximum number of spans drops down to 3. 429 For ASE the only link-dependent information needed by the routing 430 algorithm is the noise of the link, denoted as link-noise, which is 431 the sum of the noise of all spans on the link. Hence the constraint 432 on ASE becomes that the aggregate noise of the transparent segment 433 which is the sum of the link-noise of all links can not exceed 434 P/SNRmin. 436 4.4 Approximating The Effects Of Some Other Impairment Constraints 438 There are a number of other impairment constraints that we believe 439 could be approximated with a domain-wide margin on the OSNR, plus in 440 some cases a constraint on the total number of networking elements 441 (OXC or OADM) along the path. Most impairments generated at OXCs or 442 OADMs, including polarization dependent loss, coherent crosstalk, 443 and effective passband width, could be dealt with using this 444 approach. In principle, impairments generated at the nodes can be 445 bounded by system engineering rules because the node elements can be 446 designed and specified in a uniform manner. This approach is not 447 feasible with PMD and noise because neither can be uniformly 448 specified. Instead, they depend on node spacing and the 449 characteristics of the installed fiber plant, neither of which are 450 likely to be under the system designer�s control. 452 Examples of the constraints we propose to approximate with a domain- 453 wide margin are given in the remaining paragraphs in this section. 454 It should be kept in mind that as optical transport technology 455 evolves it may become necessary to include some of these impairments 456 explicitly in the routing process. Other impairments not mentioned 457 here at all may also become sufficiently important to require 458 incorporation either explicitly or via a domain-wide margin. 460 Other Polarization Dependent Impairments Other polarization- 461 dependent effects besides PMD influence system performance. For 462 example, many components have polarization-dependent loss (PDL) 463 [Ramaswami98], which accumulates in a system with many components on 464 Impairments And Other Constraints September 2002 465 On Optical Layer Routing 467 the transmission path. The state of polarization fluctuates with 468 time and its distribution is very important also. It is generally 469 required to maintain the total PDL on the path to be within some 470 acceptable limit, potentially by using some compensation technology 471 for relatively long transmission systems, plus a small built-in 472 margin in OSNR. Since the total PDL increases with the number of 473 components in the data path, it must be taken into account by the 474 system vendor when determining the maximum allowable number of 475 spans. 477 Chromatic Dispersion In general this impairment can be adequately 478 (but not optimally) compensated for on a per-link basis, and/or at 479 system initial setup time. Today most deployed compensation devices 480 are based on DCF (Dispersion Compensation Fiber). DCF provides per 481 fiber compensation by means of a spool of fiber with a CD coefficient 482 opposite to the fiber. Due to the imperfect matching between the CD 483 slope of the fiber and the DCF some lambdas can be over compensated 484 while others can be under compensated. Moreover DCF modules may only 485 be available in fixed lengths of compensating fiber; this means that 486 sometimes it is impossible to find a DCF module that exactly 487 compensates the CD introduced by the fiber. These effects introduce 488 what is known as residual CD. Residual CD varies with the frequency 489 of the wavelength. Knowing the characteristics of both of the fiber 490 and the DCF modules along the path, this can be calculated with a 491 sufficient degree of precision. However this is a very challenging 492 task. In fact the per-wavelength residual dispersion needs to be 493 combined with other information in the system (e.g. types fibers to 494 figure out the amount of nonlinearities) to obtain the net effect of 495 CD either by simulation or by some analytical approximation. It 496 appears that the routing/control plane should not be burdened by such 497 a large set of information while it can be handled at the system 498 design level. Therefore it will be assumed until proven otherwise 499 that residual dispersion should not be reported. For high bit rates, 500 dynamic dispersion compensation may be required at the receiver to 501 clean up any residual dispersion. 503 Crosstalk Optical crosstalk refers to the effect of other signals on 504 the desired signal. It includes both coherent (i.e. intrachannel) 505 crosstalk and incoherent (i.e. interchannel) crosstalk. Main 506 contributors of crosstalk are the OADM and OXC sites that use a DWDM 507 multiplexer/demultiplexer (MUX/DEMUX) pair. For a relatively sparse 508 network where the number of OADM/OXC nodes on a path is low, 509 crosstalk can be treated with a low margin in OSNR without being a 510 binding constraint. But for some relatively dense networks where 511 crosstalk might become a binding constraint, one needs to propagate 512 the per-link crosstalk information to make sure that the end-to-end 513 Impairments And Other Constraints September 2002 514 On Optical Layer Routing 516 path crosstalk which is the sum of the crosstalks on all the 517 corresponding links to be within some limit, e.g. -25dB threshold 518 with 1dB penalty ([Goldstein94]). Another way to treat it without 519 having to propagate per-link crosstalk information is to have the 520 system evaluate what the maximum number of OADM/OXC nodes that has a 521 MUX/DEMUX pair for the worst route in the transparent domain for a 522 low built-in margin. The latter one should work well where all the 523 OXC/OADM nodes have similar level of crosstalk. 525 Effective Passband As more and more DWDM components are cascaded, 526 the effective passband narrows. The number of filters along the 527 link, their passband width and their shape will determine the end- 528 to-end effective passband. In general, this is a system design 529 issue, i.e., the system is designed with certain maximum bit rate 530 using the proper modulation format and filter spacing. For linear 531 systems, the filter effect can be turned into a constraint on the 532 maximum number of narrow filters with the condition that filters in 533 the systems are at least as wide as the one in the receiver. 534 Because traffic at lower bit rates can tolerate a narrower passband, 535 the maximum allowable number of narrow filters will increase as the 536 bit rate decreases. 538 Nonlinear Impairments It seems unlikely that these can be dealt with 539 explicitly in a routing algorithm because they lead to constraints 540 that can couple routes together and lead to complex dependencies, 541 e.g. on the order in which specific fiber types are traversed 542 [Kaminow97]. Note that different fiber types (standard single mode 543 fiber, dispersion shifted fiber, dispersion compensated fiber, etc.) 544 have very different effects from nonlinear impairments. A full 545 treatment of the nonlinear constraints would likely require very 546 detailed knowledge of the physical infrastructure, including 547 measured dispersion values for each span, fiber core area and 548 composition, as well as knowledge of subsystem details such as 549 dispersion compensation technology. This information would need to 550 be combined with knowledge of the current loading of optical signals 551 on the links of interest to determine the level of nonlinear 552 impairment. Alternatively, one could assume that nonlinear 553 impairments are bounded and result in X dB margin in the required 554 OSNR level for a given bit rate, where X for performance reasons 555 would be limited to 1 or 2 dB, consequently setting a limit on the 556 maximum number of spans. For the approach described here to be 557 useful, it is desirable for this span length limit to be longer than 558 that imposed by the constraints which can be treated explicitly. 559 When designing a DWDM transport system, there are tradeoffs between 560 signal power launched at the transmitter, span length, and nonlinear 561 effects on BER that need to be considered jointly. Here, we assume 562 that an X dB margin is obtained after the transport system has been 563 designed with a fixed signal power and maximum span length for a 564 given bit rate. Note that OTSs can be designed in very different 565 ways, in linear, pseudo-linear, or nonlinear environments. The X-dB 566 Impairments And Other Constraints September 2002 567 On Optical Layer Routing 569 margin approach may be valid for some but not for others. However, 570 it is likely that there is an advantage in designing systems that 571 are less aggressive with respect to nonlinearities, and therefore 572 somewhat sub-optimal, in exchange for improved scalability, 573 simplicity and flexibility in routing and control plane design. 575 4.5 Other Impairment Considerations 577 There are many other types of impairments that can degrade 578 performance. In this section we briefly mention one other type of 579 impairment, which we propose be dealt with by either by the system 580 designer or by the transmission engineers at the time the system is 581 installed. If dealt with successfully in this manner they should not 582 need to be considered in the dynamic routing process. 584 Gain Nonuniformity and Gain Transients For simple noise estimates to 585 be of use, the amplifiers must be gain-flattened and must have 586 automatic gain control (AGC). Furthermore, each link should have 587 dynamic gain equalization (DGE) to optimize power levels each time 588 wavelengths are added or dropped. Variable optical attenuators on 589 the output ports of an OXC or OADM can be used for this purpose, and 590 in-line devices are starting to become commercially available. 591 Optical channel monitors are also required to provide feedback to 592 the DGEs. AGC must be done rapidly if signal degradation after a 593 protection switch or link failure is to be avoided. 595 Note that the impairments considered here are treated more or less 596 independently. By considering them jointly and varying the tradeoffs 597 between the effects from different components may allow more routes 598 to be feasible. If that is desirable or the system is designed such 599 that certain impairments (e.g. nonlinearities) need to be considered 600 by a centralized process, then distributed routing is not the one to 601 use. 603 4.6 An Alternative Approach - Using Maximum Distance As The only 604 Constraint 606 Today, carriers often use maximum distance to engineer point-to- 607 point OTS systems given a fixed per-span length based on the OSNR 608 constraint for a given bit rate. They may desire to keep the same 609 engineering rule when they move to all-optical networks. Here, we 610 discuss the assumptions that need to be satisfied to keep this 611 approach viable and how to treat the network elements between two 612 adjacent links. 614 Impairments And Other Constraints September 2002 615 On Optical Layer Routing 617 In order to use the maximum distance for a given bit rate to meet an 618 OSNR constraint as the only binding constraint, the operators need 619 to satisfy the following constraints in their all-optical networks: 621 - All the other non-OSNR constraints described in the previous 622 subsections are not binding factors as long as the maximum 623 distance constraint is met. 624 - Specifically for PMD, this means that the whole all-optical 625 network is built on top of sufficiently low-PMD fiber such that 626 the upper bound on the mean aggregate path DGD is always 627 satisfied for any path that does not exceed the maximum 628 distance, or PMD compensation devices might be used for routes 629 with high-PMD fibers. 630 - In terms of the ASE/OSNR constraint, in order to convert the ASE 631 constraint into a distance constraint directly, the network 632 needs to have a fixed fiber distance D for each span (so that 633 ASE can be directly mapped by the gain of the amplifier which 634 equals to the loss of the previous fiber span), e.g., 80km 635 spacing which is commonly chosen by carriers. However, when 636 spans have variable lengths, certain adjustment and compromise 637 need to be made in order to avoid treating ASE explicitly as in 638 section 4.3. These include: 1) If a span is shorter than a 639 typical span length D, unless certain mechanism is built in the 640 OTS to take advantages of shorter spans, it needs to be treated 641 as a span of length D instead of with its real length. 2) When 642 there are spans that are longer than D, it means that paths with 643 these longer spans would have higher average span loss. In 644 general, the maximum system reach decreases when the average 645 span loss increases. Thus, in order to accommodate longer spans 646 in the network, the maximum distance upper bound has to be set 647 with respect to the average span loss of the worst path in the 648 network. This sub-optimality may be acceptable for some networks 649 if the variance is not too large, but may be too conservative 650 for others. 652 If these assumptions are satisfied, the second issue we need to 653 address is how to treat a transparent network element (e.g., MEMS- 654 based switch) between two adjacent links in terms of a distance 655 constraint since it also introduces an insertion loss. If the 656 network element cannot somehow compensate for this OSNR degradation, 657 one approach is to convert each network element into an equivalent 658 length of fiber based on its loss/ASE contribution. Hence, in 659 general, introducing a set of transparent network elements would 660 Impairments And Other Constraints September 2002 661 On Optical Layer Routing 663 effectively result in reducing the overall actual transmission 664 distance between the OEO edges. 666 With this approach, the link-specific state information is link- 667 distance, the length of a link. It equals to the distance sum of all 668 fiber spans on the link and the equivalent length of fiber for the 669 network element(s) on the link. The constraint is that the sum of 670 all the link-distance over all links of a path should be less than 671 the maximum-path-distance, the upper bound of all paths. 673 4.7 Other Considerations 675 Routing in an all-optical network without wavelength conversion 676 raises several additional issues: 678 - Since the route selected must have the chosen wavelength 679 available on all links, this information needs to be considered 680 in the routing process. This is discussed in [Chaudhuri00], 681 where it is concluded that advertising detailed wavelength 682 availabilities on each link is not likely to scale. Instead 683 they propose an alternative method which probes along a chosen 684 path to determine which wavelengths (if any) are available. 685 This would require a significant addition to the routing logic 686 normally used in OSPF. Others have proposed simultaneously 687 probing along multiple paths. 689 - Choosing a path first and then a wavelength along the path is 690 known to give adequate results in simple topologies such as 691 rings and trees ([Yates99]). This does not appear to be true in 692 large mesh networks under realistic provisioning scenarios, 693 however. Instead significantly better results are achieved if 694 wavelength and route are chosen simultaneously ([Strand01b]). 695 This approach would however also have a significant effect on 696 OSPF. 698 4.8 Implications For Routing and Control Plane Design 700 If distributed routing is desired, additional state information will 701 be required by the routing to deal with the impairments described in 702 Sections 4.2 - 4.4: 704 - As mentioned earlier, an operator who wants to avoid having to 705 provide impairment-related parameters to the control plane may 706 elect not to deal with them at the routing level, instead 707 treating them at the system design and planning level if that is 708 a viable approach for their network. In this approach the 709 operator can pre-qualify all or a set of feasible end-to-end 710 Impairments And Other Constraints September 2002 711 On Optical Layer Routing 713 optical paths through the domain of transparency for each bit 714 rate. This approach may work well with relatively small and 715 sparse networks, but it may not be scalable for large and dense 716 networks where the number of feasible paths can be very large. 718 - If the optical paths are not pre-qualified, additional link- 719 specific state information will be required by the routing 720 algorithm for each type of impairment that has the potential of 721 being limiting for some routes. Note that for one operator, PMD 722 might be the only limiting constraint while for another, ASE 723 might be the only one, or it could be both plus some other 724 constraints considered in this document. Some networks might not 725 be limited by any of these constraints. 727 - For an operator needing to deal explicitly with these 728 constraints, the link-dependent information identified above for 729 PMD is link-PMD-square which is the square of the total PMD on a 730 link. For ASE the link-dependent information identified is link- 731 noise which is the total noise on a link. Other link-dependent 732 information includes link-span-length which is the total number 733 of spans on a link, link-crosstalk or OADM-OXC-number which is 734 the total crosstalk or the number of OADM/OXC nodes on a link, 735 respectively, and filter-number which is the number of narrow 736 filters on a link. When the alternative distance-only approach 737 is chosen, the link-specific information is link-distance. 739 - In addition to the link-specific information, bounds on each of 740 the impairments need to be quantified. Since these bounds are 741 determined by the system designer's impairment allocations, 742 these will be system dependent. For PMD, the constraint is that 743 the sum of the link-PMD-square of all links on the transparent 744 segment is less than the square of (a/B) where B is the bit 745 rate. Hence, the required information is the parameter "a". For 746 ASE, the constraint is that the sum of the link-noise of all 747 links is no larger than P/SNRmin. Thus, the information needed 748 include the launch power P and OSNR requirement SNRmin. The 749 minimum acceptable OSNR, in turn, depends on the strength of the 750 FEC being used and the margins reserved for other types of 751 impairments. Other bounds include the maximum span length of the 752 transmission system, the maximum path crosstalk or the maximum 753 number of OADM/OXC nodes, and the maximum number of narrow 754 filters, all are bit rate dependent. With the alternative 755 distance-only approach, the upper bound is the maximum-path- 756 distance. In single-vendor "islands" some of these parameters 757 may be available in a local or EMS database and would not need 758 to be advertised 760 - It is likely that the physical layer parameters do not change 761 value rapidly and could be stored in some database; however 762 these are physical layer parameters that today are frequently 763 Impairments And Other Constraints September 2002 764 On Optical Layer Routing 766 not known at the granularity required. If the ingress node of a 767 lightpath does path selection these parameters would need to be 768 available at this node. 770 - The specific constraints required in a given situation will 771 depend on the design and engineering of the domain of 772 transparency; for example it will be essential to know whether 773 chromatic dispersion has been dealt with on a per-link basis, 774 and whether the domain is operating in a linear or nonlinear 775 regime. 777 - As optical transport technology evolves, the set of constraints 778 that will need to be considered either explicitly or via a 779 domain-wide margin may change. The routing and control plane 780 design should therefore be as open as possible, allowing 781 parameters to be included as necessary. 783 - In the absence of wavelength conversion, the necessity of 784 finding a single wavelength that is available on all links 785 introduces the need to either advertise detailed information on 786 wavelength availability, which probably doesn't scale, or have 787 some mechanism for probing potential routes with or without 788 crankback to determine wavelength availability. Choosing the 789 route first, and then the wavelength, may not yield acceptable 790 utilization levels in mesh-type networks. 792 5. More Complex Networks 794 Mixing optical equipment in a single domain of transparency that has 795 not been explicitly designed to interwork is beyond the scope of 796 this document. This includes most multi-vendor all-optical networks. 798 An optical network composed of multiple domains of transparency 799 optically isolated from each other by O/E/O devices (transponders) 800 is more plausible. A network composed of both "opaque" (optically 801 isolated) OLXC's and one or more all-optical "islands" isolated by 802 transponders is of particular interest because this is most likely 803 how all-optical technologies (such as that described in Sec. 2) are 804 going to be introduced. (We use the term "island" in this discussion 805 rather than a term like "domain" or "area" because these terms are 806 associated with specific approaches like BGP or OSPF.) 808 We consider the complexities raised by these alternatives now. 810 The first requirement for routing in a multi-island network is that 811 the routing process needs to know the extent of each island. There 812 are several reasons for this: 814 Impairments And Other Constraints September 2002 815 On Optical Layer Routing 817 - When entering or leaving an all-optical island, the regeneration 818 process cleans up the optical impairments discussed in Sec. 3. 819 - Each all-optical island may have its own bounds on each 820 impairment. 821 - The routing process needs to be sensitive to the costs 822 associated with "island-hopping". 824 This last point needs elaboration. It is extremely important to 825 realize that, at least in the short to intermediate term, the 826 resources committed by a single routing decision can be very 827 significant: The equipment tied up by a single coast-to-coast OC-192 828 can easily have a first cost of $10**6, and the holding times on a 829 circuit once established is likely to be measured in months. 830 Carriers will expect the routing algorithms used to be sensitive to 831 these costs. Simplistic measures of cost such as the number of 832 "hops" are not likely to be acceptable. 834 Taking the case of an all-optical island consisting of an "ultra 835 long-haul" system like that in Fig. 3-1 embedded in an OEO network 836 of electrical fabric OLXC's as an example: It is likely that the ULH 837 system will be relatively expensive for short hops but relatively 838 economical for longer distances. It is therefore likely to be 839 deployed as a sort of "express backbone". In this scenario a carrier 840 is likely to expect the routing algorithm to balance OEO costs 841 against the additional costs associated with ULH technology and 842 route circuitously to make maximum use of the backbone where 843 appropriate. Note that the metrics used to do this must be 844 consistent throughout the routing domain if this expectation is to 845 be met. 847 The first-order implications for GMPLS seem to be: 848 - Information about island boundaries needs to be advertised. 849 - The routing algorithm needs to be sensitive to island 850 transitions and to the connectivity limitations and impairment 851 constraints particular to each island. 852 - The cost function used in routing must allow the balancing of 853 transponder costs, OXC and OADM costs, and line haul costs 854 across the entire routing domain. 856 Several distributed approaches to multi-island routing seem worth 857 investigating: 858 - Advertise the internal topology and constraints of each island 859 globally; let the ingress node compute an end-to-end strict 860 explicit route sensitive to all constraints and wavelength 861 availabilities. In this approach the routing algorithm used by 862 the ingress node must be able to deal with the details of 863 routing within each island. 864 - Have the EMS or control plane of each island determine and 865 advertise the connectivity between its boundary nodes together 866 with additional information such as costs and the bit rates and 867 Impairments And Other Constraints September 2002 868 On Optical Layer Routing 870 formats supported. As the spare capacity situation changes, 871 updates would be advertised. In this approach impairment 872 constraints are handled within each island and impairment- 873 related parameters need not be advertised outside of the island. 874 The ingress node would then do a loose explicit route and leave 875 the routing and wavelength selection within each island to the 876 island. 877 - Have the ingress node send out probes or queries to nearby 878 gateway nodes or to an NMS to get routing guidance. 880 6. Diversity 882 6.1 Background On Diversity 884 "Diversity" is a relationship between lightpaths. Two lightpaths are 885 said to be diverse if they have no single point of failure. In 886 traditional telephony the dominant transport failure mode is a 887 failure in the interoffice plant, such as a fiber cut inflicted by a 888 backhoe. 890 Why is diversity a unique problem that needs to be considered for 891 optical networks? So far, data network operators have relied on 892 their private line providers to ensure diversity and so have not had 893 to deal directly with the problem. GMPLS makes the complexities 894 handled by the private line provisioning process, including 895 diversity, part of the common control plane and so visible to all. 897 To determine whether two lightpath routings are diverse it is 898 necessary to identify single points of failure in the interoffice 899 plant. To do so we will use the following terms: A fiber cable is a 900 uniform group of fibers contained in a sheath. An Optical Transport 901 System will occupy fibers in a sequence of fiber cables. Each fiber 902 cable will be placed in a sequence of conduits - buried honeycomb 903 structures through which fiber cables may be pulled - or buried in a 904 right of way (ROW). A ROW is land in which the network operator has 905 the right to install his conduit or fiber cable. It is worth noting 906 that for economic reasons, ROW�s are frequently obtained from 907 railroads, pipeline companies, or thruways. It is frequently the 908 case that several carriers may lease ROW from the same source; this 909 makes it common to have a number of carriers� fiber cables in close 910 proximity to each other. Similarly, in a metropolitan network, 911 several carriers might be leasing duct space in the same RBOC 912 conduit. There are also "carrier's carriers" - optical networks 913 which provide fibers to multiple carriers, all of whom could be 914 affected by a single failure in the "carrier's carrier" network. 916 Impairments And Other Constraints September 2002 917 On Optical Layer Routing 919 In a typical intercity facility network there might be on the order 920 of 100 offices that are candidates for OLXC�s. To represent the 921 inter-office fiber network accurately a network with an order of 922 magnitude more nodes is required. In addition to Optical Amplifier 923 (OA) sites, these additional nodes include: 924 - Places where fiber cables enter/leave a conduit or right of way; 925 - Locations where fiber cables cross; 926 Locations where fiber splices are used to interchange fibers 927 between fiber cables. 928 An example of the first might be: 929 A B 930 A-------------B \ / 931 \ / 932 X-----Y 933 / \ 934 C-------------D / \ 935 C D 937 (a) Fiber Cable Topology (b) Right-Of-Way/Conduit Topology 939 Figure 6-1: Fiber Cable vs. ROW Topologies 941 Here the A-B fiber cable would be physically routed A-X-Y-B and the 942 C-D cable would be physically routed C-X-Y-D. This topology might 943 arise because of some physical bottleneck: X-Y might be the Lincoln 944 Tunnel, for example, or the Bay Bridge. 946 Fiber route crossing (the second case) is really a special case of 947 this, where X and Y coincide. In this case the crossing point may 948 not even be a manhole; the fiber routes might just be buried at 949 different depths. 951 Fiber splicing (the third case) often occurs when a major fiber 952 route passes near to a small office. To avoid the expense and 953 additional transmission loss only a small number of fibers are 954 spliced out of the major route into a smaller route going to the 955 small office. This might well occur in a manhole or hut. An 956 example is shown in Fig. 6-2(a), where A-X-B is the major route, X 957 the manhole, and C the smaller office. The actual fiber topology 958 would then look like Fig. 6-2(b), where there would typically be 959 many more A-B fibers than A-C or C-B fibers, and where A-C and C-B 960 might have different numbers of fibers. (One of the latter might 961 even be missing.) 962 Impairments And Other Constraints September 2002 963 On Optical Layer Routing 965 C C 966 | / \ 967 | / \ 968 | / \ 969 A------X------B A---------------B 971 (a) Fiber Cable Topology (b) Fiber Topology 973 Figure 6-2. Fiber Cable vs Fiber Topologies 975 The imminent deployment of ultra-long (>1000 km) Optical Transport 976 Systems introduces a further complexity: Two OTS's could interact a 977 number of times. To make up a hypothetical example: A New York - 978 Atlanta OTS and a Philadelphia - Orlando OTS might ride on the same 979 right of way for x miles in Maryland and then again for y miles in 980 Georgia. They might also cross at Raleigh or some other intermediate 981 node without sharing right of way. 983 Diversity is often equated to routing two lightpaths between a 984 single pair of points, or different pairs of points so that no 985 single route failure will disrupt them both. This is too simplistic, 986 for a number of reasons: 988 - A sophisticated client of an optical network will want to derive 989 diversity needs from his/her end customers' availability 990 requirements. These often lead to more complex diversity 991 requirements than simply providing diversity between two 992 lightpaths. For example, a common requirement is that no single 993 failure should isolate a node or nodes. If a node A has single 994 lightpaths to nodes B and C, this requires A-B and A-C to be 995 diverse. In real applications, a large data network with N 996 lightpaths between its routers might describe their needs in an 997 NxN matrix, where (i,j) defines whether lightpaths i and j must 998 be diverse. 1000 - Two circuits that might be considered diverse for one 1001 application might not be considered diverse for in another 1002 situation. Diversity is usually thought of as a reaction to 1003 interoffice route failures. High reliability applications may 1004 require other types of failures to be taken into account. Some 1005 examples: 1006 o Office Outages: Although less frequent than route failures, 1007 fires, power outages, and floods do occur. Many network 1008 Impairments And Other Constraints September 2002 1009 On Optical Layer Routing 1011 managers require that diverse routes have no (intermediate) 1012 nodes in common. In other cases an intermediate node might 1013 be acceptable as long as there is power diversity within 1014 the office. 1015 o Shared Rings: Many applications are willing to allow 1016 "diverse" circuits to share a SONET ring-protected link; 1017 presumably they would allow the same for optical layer 1018 rings. 1019 o Disasters: Earthquakes and floods can cause failures over 1020 an extended area. Defense Department circuits might need 1021 to be routed with nuclear damage radii taken into account. 1022 - Conversely, some networks may be willing to take somewhat larger 1023 risks. Taking route failures as an example: Such a network 1024 might be willing to consider two fiber cables in heavy duty 1025 concrete conduit as having a low enough chance of simultaneous 1026 failure to be considered "diverse". They might also be willing 1027 to view two fiber cables buried on opposite sides of a railroad 1028 track as being diverse because there is minimal danger of a 1029 single backhoe disrupting them both even though a bad train 1030 wreck might jeopardize them both. A network seeking N mutually 1031 diverse paths from an office with less than N diverse ROW�s will 1032 need to live with some level of compromise in the immediate 1033 vicinity of the office. 1035 These considerations strongly suggest that the routing algorithm 1036 should be sensitive to the types of threat considered unacceptable 1037 by the requester. Note that the impairment constraints described in 1038 the previous section may eliminate some of the long circuitous 1039 routes sometimes needed to provide diversity. This would make it 1040 harder to find many diverse paths through an all-optical network 1041 than an opaque one. 1043 [Chaudhuri00] introduced the term "Shared Risk Link Group" (SRLG) to 1044 describe the relationship between two non-diverse links. The above 1045 discussion suggests that an SRLG should be characterized by 2 1046 parameters: 1047 - Type of Compromise: Examples would be shared fiber cable, shared 1048 conduit, shared ROW, shared optical ring, shared office without 1049 power sharing, etc.) 1050 - Extent of Compromise: For compromised outside plant, this would 1051 be the length of the sharing. 1052 A CSPF algorithm could then penalize a diversity compromise by an 1053 amount dependent on these two parameters. 1055 Impairments And Other Constraints September 2002 1056 On Optical Layer Routing 1058 Two links could be related by many SRLG's (AT&T's experience 1059 indicates that a link may belong to over 100 SRLG's, each 1060 corresponding to a separate fiber group. Each SRLG might relate a 1061 single link to many other links. For the optical layer, similar 1062 situations can be expected where a link is an ultra-long OTS). 1064 The mapping between links and different types of SRLG�s is in 1065 general defined by network operators based on the definition of each 1066 SRLG type. Since SRLG information is not yet ready to be 1067 discoverable by a network element and does not change dynamically, 1068 it need not be advertised with other resource availability 1069 information by network elements. It could be configured in some 1070 central database and be distributed to or retrieved by the nodes, or 1071 advertised by network elements at the topology discovery stage. 1073 6.2 Implications For Routing 1075 Dealing with diversity is an unavoidable requirement for routing in 1076 the optical layer. It requires dealing with constraints in the 1077 routing process but most importantly requires additional state 1078 information � the SRLG relationships and also the routings of any 1079 existing circuits from the new circuit is to be diverse � to be 1080 available to the routing process. 1082 At present SRLG information cannot be self-discovered. Indeed, in a 1083 large network it is very difficult to maintain accurate SRLG 1084 information. The problem becomes particularly daunting whenever 1085 multiple administrative domains are involved, for instance after the 1086 acquisition of one network by another, because there normally is a 1087 likelihood that there are diversity violations between the domains. 1088 It is very unlikely that diversity relationships between carriers 1089 will be known any time in the near future. 1091 Considerable variation in what different customers will mean by 1092 acceptable diversity should be anticipated. Consequently we suggest 1093 that an SRLG should be defined as follows: (i) It is a relationship 1094 between two or more links, and (ii) it is characterized by two 1095 parameters, the type of compromise (shared conduit, shared ROW, 1096 shared optical ring, etc.) and the extent of the compromise (e.g., 1097 the number of miles over which the compromise persisted). This will 1098 allow the SRLG�s appropriate to a particular routing request to be 1099 easily identified. 1101 7. Security Considerations 1102 Impairments And Other Constraints September 2002 1103 On Optical Layer Routing 1105 We are assuming OEO interfaces to the domain(s) covered by our 1106 discussion (see, e.g., Sec. 4.1 above). If this assumption were to 1107 be relaxed and externally generated optical signals allowed into the 1108 domain, network security issues would arise. Specifically, 1109 unauthorized usage in the form of signals at improper wavelengths or 1110 with power levels or impairments inconsistent with those assumed by 1111 the domain would be possible. With OEO interfaces, these types of 1112 layer one threats should be controllable. 1114 A key layer one security issue is resilience in the face of physical 1115 attack. Diversity, as describe in Sec. 6, is a part of the 1116 solution. However, it is ineffective if there is not sufficient 1117 spare capacity available to make the network whole after an attack. 1118 Several major related issues are: 1119 - Defining the threat: If, for example, an electro-magnetic 1120 interference (EMI) burst is an in-scope threat, then (in the 1121 terminology of Sec. 6) all of the links sufficiently close 1122 together to be disrupted by such a burst must be included in a 1123 single SRLG. Similarly for other threats: For each in-scope 1124 threat, SRLG�s must be defined so that all links vulnerable to a 1125 single incident of the threat must be grouped together in a 1126 single SRLG. 1127 - Allocating responsibility for responding to a layer one failure 1128 between the various layers (especially the optical and IP 1129 layers): This must be clearly specified to avoid churning and 1130 unnecessary service interruptions. 1132 The whole proposed process depends on the integrity of the 1133 impairment characterization information (PMD parameters, etc.) and 1134 also the SRLG definitions. Security of this information, both when 1135 stored and when distributed, is essential. 1137 This document does not address control plane issues, and so control- 1138 plane security is out of scope. 1140 8. Acknowledgments 1142 This document has benefited from discussions with Michael Eiselt, 1143 Jonathan Lang, Mark Shtaif, Jennifer Yates, Dongmei Wang, Guangzhi 1144 Li, Robert Doverspike, Albert Greenberg, Jim Maloney, John Jacob, 1145 Katie Hall, Diego Caviglia, D. Papadimitriou, O. Audouin, J. P. 1146 Faure, L. Noirie, and with our OIF colleagues. 1148 9. References 1150 9.1 Normative References 1151 Impairments And Other Constraints September 2002 1152 On Optical Layer Routing 1154 [Chaudhuri00] Chaudhuri, S., Hjalmtysson, G., and Yates, J., 1155 "Control of Lightpaths in an Optical Network", Work in Progress, 1156 draft-chaudhuri-ip-olxc-control-00.txt. 1158 [Goldstein94] Goldstein, E. L., Eskildsen, L., and Elrefaie, A. F., 1159 Performance Implications of Component Crosstalk in Transparent 1160 Lightwave Networks", IEEE Photonics Technology Letters, Vol.6, No.5, 1161 May 1994. 1163 [ITU] ITU-T Doc. G.663, Optical Fibers and Amplifiers, Section 1164 II.4.1.2. 1166 [Kaminow97] Kaminow, I. P. and Koch, T. L., editors, Optical Fiber 1167 Telecommunications IIIA, Academic Press, 1997. 1169 [Strand01] J. Strand, A. Chiu, and R. Tkach, "Issues for Routing in 1170 the Optical Layer", IEEE Communications Magazine, Feb. 2001, vol. 39 1171 No. 2, pp. 81-88; also see "Unique Features and Requirements for The 1172 Optical Layer Control Plane", Internet Draft, draft-chiu-strand- 1173 unique-olcp-01.txt, work in progress, November 2000. 1175 [Strand01b] J. Strand, R. Doverspike, and G. Li, "Importance of 1176 Wavelength Conversion In An Optical Network", Optical Networks 1177 Magazine, May/June 2001, pp. 33-44. 1179 [Yates99] Yates, J. M., Rumsewicz, M. P. and Lacey, J. P. R., 1180 "Wavelength Converters in Dynamically-Reconfigurable WDM Networks", 1181 IEEE Communications Surveys, 2Q1999 (online at 1182 www.comsoc.org/pubs/surveys/2q99issue/yates.html). 1184 9.2 Informative References 1186 [Awduche99] Awduche, D. O., Rekhter, Y., Drake, J., and Coltun, R., 1187 "Multi-Protocol Lambda Switching: Combining MPLS Traffic Engineering 1188 Control With Optical Crossconnects", Work in Progress, draft- 1189 awduche-mpls-te-optical-01.txt. 1191 [Bra96] Bradner, S., "The Internet Standards Process -- Revision 3," 1192 BCP 9, RFC 2026, October 1996. 1194 [CBD00] Ceuppens, L., Blumenthal, D., Drake, J., Chrostowski, J., 1195 Edwards, W., "Performance Monitoring in Photonic Networks in Support 1196 of MPL(ambda)S", Internet draft, work in progress, March 2000. 1198 Impairments And Other Constraints September 2002 1199 On Optical Layer Routing 1201 [Doverspike00] Doverspike, R. and Yates, J., "Challenges For MPLS in 1202 Optical Network Restoration", IEEE Communication Magazine, February, 1203 2001. 1205 [Gerstel 2000] O. Gorstel, "Optical Layer Signaling: How Much Is 1206 Really Needed?" IEEE Communications Magazine, vol. 38 no. 10, Oct. 1207 2000, pp. 154-160 1209 [KRB01a] Kompella, K., et.al., "IS-IS extensions in support of 1210 Generalized MPLS," Internet Draft, draft-ietf-gmpls- extensions- 1211 01.txt, work in progress, 2001. 1213 [KRB01b] Kompella, K., et. al., "OSPF extensions in support of 1214 Generalized MPLS," Internet draft, draft-ospf-generalized- mpls- 1215 00.txt, work in progress, March 2001. 1217 [Moy98] Moy, John T., OSPF: Anatomy of an Internet Routing Protocol, 1218 Addison-Wesley, 1998. 1220 [Passmore01] Passmore, D., "Managing Fatter Pipes," Business 1221 Communications Review, August 2001, pp. 20-21. 1223 [Ramaswami98] Ramaswami, R. and Sivarajan, K. N., Optical Networks: 1224 A Practical Perspective, Morgan Kaufmann Publishers, 1998. 1226 [Tkach98] Tkach, R., Goldstein, E., Nagel, J., and Strand, J., 1227 "Fundamental Limits of Optical Transparency", Optical Fiber 1228 Communication Conf., Feb. 1998, pp. 161-162. 1230 10. Contributing Authors 1232 This document was a collective work of a number of people. The text 1233 and content of this document was contributed by the editors and the 1234 co-authors listed below. 1236 Ayan Banerjee 1237 Calient Networks 1238 5853 Rue Ferrari 1239 San Jose, CA 95138 1240 Email: abanerjee@calient.net 1242 Dan Blumenthal 1243 Calient Networks 1244 Impairments And Other Constraints September 2002 1245 On Optical Layer Routing 1247 5853 Rue Ferrari 1248 San Jose, CA 95138 1249 Email: dblumenthal@calient.net 1251 John Drake 1252 Calient Networks 1253 5853 Rue Ferrari 1254 San Jose, CA 95138 1255 Email: jdrake@calient.net 1257 Andre Fredette 1258 Hatteras Networks 1259 PO Box 110025 1260 Research Triangle Park, NC 27709 1261 Email: afredette@hatterasnetworks.com 1263 Nan Froberg 1264 PhotonEx Corporation 1265 200 Metrowest Technology Dr. 1266 Maynard, MA 01754 1267 Email: nfroberg@photonex.com 1269 Taha Landolsi 1270 WorldCom, Inc. 1271 2400 North Glenville Drive 1272 Richardson, TX 75082 1273 Email: taha.landolsi@wcom.com 1275 James V. Luciani 1276 900 Chelmsford St. 1277 Lowell, MA 01851 1278 Email: james_luciani@mindspring.com 1280 Robert Tkach 1281 Celion Networks 1282 1 Sheila Dr., Suite 2 1283 Tinton Falls, NJ 07724 1284 Email: bob.tkach@celion.com 1286 Yong Xue 1287 WorldCom, Inc. 1288 22001 Loudoun County Parkway 1289 Ashburn, VA 20147 1290 Email: yxue@cox.com 1292 11. Editors� Addresses 1294 Angela Chiu 1295 Impairments And Other Constraints September 2002 1296 On Optical Layer Routing 1298 Celion Networks 1299 1 Sheila Dr., Suite 2 1300 Tinton Falls, NJ 07724 1301 Phone:(732) 747-9987 1302 Email: alchiu@ieee.org 1304 John Strand 1305 AT&T Labs 1306 200 Laurel Ave., Rm A5-1D33 1307 Middletown, NJ 07748 1308 Phone:(732) 420-9036 1309 Email: jls@research.att.com